NCEP Reanalysis data problem

[brett_kuyper]

Problem:

I am trying to run the Hysplit model to generate or calculate back trajectories in R using David Carslaw's OPENAIR methods (https://gist.github.com/davidcarslaw/c6 ... 7796e4bdf5). For this I have been using the NCEP/NCAR Reanalysis Data Archive obtained from ftp://arlftp.arlhq.noaa.gov/archives/reanalysis/RP/.
The method runs well and the back trajectories are calculated. However, the model fails when either RP201707.gbl or RP201708.gbl are used as the met files.
The error returned:
"Error in read.table(file, header = FALSE, skip = 7) :
no lines available in input. "
The met_data for months on either side of July/August 2017 computes, calculates and binds without a problem. This suggests to me that there is problem with these met data. It is possible that in the downloading something has gone awry, although I have tried repeatedly. Please is there an error in these files?


Response form Barbara Stunder:

To test the ARL reanalysis data, run the trajectories that are failing on READY - https://www.ready.noaa.gov/HYSPLIT_traj.php. If successful there, then download again.

Update:

Both months appear to run on the READY website. The data has been downloaded from ftp://arlftp.arlhq.noaa.gov/archives/reanalysis/ and
ftp://arlftp.arlhq.noaa.gov/pub/archives/reanalysis/
I accidentally ran the model for the first 30 days of July, and it ran without hassle. If I try to calculate the trajectories over 31st July I get an error. The August data spits an error regardless of the days duration.

Any advice or suggestions would be most appreciated.

[barbara.stunder]

The data on READY are the same files as on the the ftp site. Try running the July 31 and an August trajectory manually without going through R. I think that error message is from R.

[brett_kuyper]

The Hysplit mode was run manually for 31 July 2017 using the RP201707.gbl data set. The following error was produced:

Calculation Started ... please be patient
*ERROR* metpos: start point not within (x,y,t) any data file
- start time before start of meteorology data
- start time after end of meteorology data
STOP 900

This error starts if the model date is set to 17 07 30 19:00. The NOAA Ready website does not have an option to calculate trajectories arriving on 31 July 2017.

The 1 August data set fails in R. This runs in manual mode without errors. On the Ready website, back trajectories calculated to arrive 1 August are calculated to midnight and do not extend into July. The rest of July and August both run well in both manual mode and in R.

Any advice would be most appreciated.

[glenn.rolph]

The 31st of July is missing from the reanalysis file. We will look into creating a new file. Thanks.

[glenn.rolph]

We updated the RP201707.gbl file and now it includes the 31st of July.

[brett_kuyper]

Thank you for following up with this and generating an updated data set.
Most appreciated